Implementing big data: 10 tips from financial service providers

A framework for deriving value from data

Many financial service providers assume that they don’t have enough information at hand. That isn’t the case: financial service providers (FSPs) often already have a wealth of high-quality information and can use data of all shapes and sizes to inform and improve their work. Yet an FSP’s data is only useful insofar as the information can be mined and analyzed to extract meaningful insights. And building the operational capacity to respond to data is equally, if not more, important. Introducing data-driven innovations is a multi-dimensional challenge for any FSP, requiring the talents and coordination of teams across the organization.

Accion’s latest Insights paper, Unlocking the Promise of (Big) Data to Promote Financial Inclusion, compiles recommendations from working with financial service providers across the globe and interviewing more than 30 industry experts. Here we share their collective knowledge for successfully implementing big data to improve their operations and accelerate financial inclusion.

implementing big data blog Unlocking the Promise of Big Data to Promote Financial Inclusion

Ten tips to follow when implementing big data

  1. Build bridges, break silos

Regardless of the scope of the project, the exercise will require coordination and collaboration among all stakeholders of the business. A good practice is to establish a working group of representatives of each business function or department (including upper management and others involved in setting the strategic direction of the organization) to see how their data management issues are interrelated.

  1. “Play with the data”

Paul Makin, co-founder of M-PESA, believes organizations should first allow their business team to play around with available data on tools such as Excel, simply “looking for interesting patterns.” The questions they begin to ask from the data should be used to inform the goals of any data project.

  1. Balance quick wins with big wins

In the short term, an organization’s data strategy and capabilities should address immediate needs to inform business decisions that require urgent analytical support. In the long term, organizations should build a strategy to acquire a unique information advantage.

  1. Do the Robot

Every human assumption needs to be picked apart to get at the logic underneath if a process is to be successfully analyzed and automated. Mapping the entire business workflow can help answer these questions and build a “data inventory.” Alex Holderness, Senior Vice President and Head of Digital at Accion in the U.S., finds it helpful to frame process-mapping conversations with staff as if you were the database. This facilitates taking a systems perspective: ask, for example, what is that information used for and how does it flow through the system? How does that information inform any decision? If we automate it, how do we get the system to learn what’s important?

  1. Make a data “wish list”

Using the process map and data inventory as working documents, have the established task force 1) identify areas for improvement in the processes, and 2) create a wish list of data points that, if available, would make decision-making and responsiveness to the customer more efficient. During this workshop, you may want to ask, “What information would you like to have, if cost were not an issue?” In addition to building capability to better leverage internal data, FSPs should consider whether to integrate supplementary data sources, including social media or government datasets, bill payment history, or call data records. Consider the potential sources of that data and the feasibility of adding them to your data inventory, in terms of both acquisition cost and ease of integration.implementing big data blog Unlocking the Promise of Big Data to Promote Financial Inclusion

  1. “Be prepared for change” – and willing to own the process

When asked for his main message to organizations looking to build data systems, Paulo Marques, a senior solutions architect at Capgemini told us: “Be prepared for change. Be prepared to encounter something every few months that will challenge your assumptions.”

Organizations must be willing to own the entire process and be receptive to change, whether for building in-house data capability or outsourcing. Amee Parbhoo, a Senior Investment Officer at Accion, and Mike Ogbalu, Divisional CEO at Interswitch, both cited FSPs expecting fintech partners to drive the change as a reason that new initiatives failed. Use the task force set up earlier in the implementation process to help develop a change management strategy and shepherd it through the organization.

Many organizations struggle with the challenge of building a new, data-driven culture that permeates up and down the ranks, but for those that succeed, senior management is key to setting and defining an innovative work culture. Identifying a key metric for success can free your organization to think creatively about problem-solving. NeoGrowth has implemented a wide range of exciting data innovations to support its growth strategy. Raju Shetty, NeoGrowth’s Chief Technology Officer, said much of this innovation was possible due to the organization’s laser focus on a single problem; how to reduce customer acquisition costs.

  1. Use common sense

FSPs need to build checks and balances into data-based decision-making processes to ensure results are backed by their real-world experience, business knowledge, and understanding of their customers. As Aneesh Varma, Aire Founder and CEO, describes his company’s central thesis, “The human mind has the best potential ability to discern creditworthiness, as it understands context really well. Aire is attempting to therefore [use data and technology to] emulate the same fundamentals that loan officers have been using for decades.”

  1. Build a team that will “chase the data”

FSPs need to develop both the analytical capacity to monitor and interpret the trends evident in their data and the operational capacity to respond to these real-time business insights.

Interviews shared that it is not necessary to hire highly credentialed analysts in the initial stages of building data capabilities. According to Quona Capital’s Ganesh Rengaswamy, a vast majority of the early work performed by data analytics teams is focused on the labor-intensive tasks of collecting and cleaning data. He recommends putting together a team that is willing to chase data, rather than a team that expects to receive pre-packaged data.

  1. Protect your customers’ data – for your own good

Jonathan Hakim of Cignifi believes it is a strategic decision to adhere to high standards of customer protection, because that is how the market is going to evolve, and it makes business sense to consider what constitutes ethical use of data.

  1. Treat partners like partners – be sensitive to their needs

As with internal change management, fintech service providers emphasized the importance of regular and clear communication for successful project management when working with vendors or other third-party partners, for example when rolling out a new data platform, product, or service.

Ensure lines of communication extend through to the team responsible for product delivery. One major African bank struggled to develop a viable mobile money platform in a key market in part because its technical service provider had subcontracted aspects of product development to another third-party vendor whom they had difficulty accessing.

Ben Knelman, Co-Founder and CEO of Juntos Finanzas, which has worked with a wide range of FSPs on customer engagement, likened managing new partnerships to the early days of a marriage: you need to be sensitive to your partner’s needs and adjust your behavior accordingly.

The last word

Invoking big data is exciting, but as Ben Knelman of Juntos noted, “Real innovation sits where it looks the least glamorous and is the most painful.” Hopefully, this guide will minimize the pain!

Sign up to receive emails